Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding
Few-shot font generation seeks to create high-quality fonts using minimal reference style images, addressing traditional font design’s labor-intensive and time-consuming nature, particularly for languages with large character sets like Chinese and Korean. Existing methods often require multi-stage t...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-02-01
|
| Series: | Applied Sciences |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2076-3417/15/3/1654 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850068507526430720 |
|---|---|
| author | Irfanullah Memon Muhammad Ammar Ul Hassan Jaeyoung Choi |
| author_facet | Irfanullah Memon Muhammad Ammar Ul Hassan Jaeyoung Choi |
| author_sort | Irfanullah Memon |
| collection | DOAJ |
| description | Few-shot font generation seeks to create high-quality fonts using minimal reference style images, addressing traditional font design’s labor-intensive and time-consuming nature, particularly for languages with large character sets like Chinese and Korean. Existing methods often require multi-stage training or predefined components, which can be time-consuming and limit generalizability. This paper introduces Patch-Font, a novel single-stage method that overcomes the limitations of prior approaches, such as multi-stage training or reliance on predefined components, by integrating a patch-based attention mechanism and a multitask encoder. Patch-Font jointly captures global style elements (e.g., overall font family characteristics) and local style details (e.g., serifs, stroke shapes), ensuring high fidelity to the target style while maintaining computational efficiency. Our approach incorporates triplet margin loss with hard positive/negative mining to disentangle style from content and a style fidelity loss to enhance local style consistency. Experiments on Korean (printed and handwritten) and Chinese fonts demonstrate that Patch-Font outperforms state-of-the-art methods in style accuracy, perceptual quality, and generation speed while generalizing robustly to unseen characters and font styles. By simplifying the font creation process and delivering high-quality results, Patch-Font represents a significant step forward in making font design more accessible and scalable for diverse languages. |
| format | Article |
| id | doaj-art-13a12585b60d450b823050814e5bc4c4 |
| institution | DOAJ |
| issn | 2076-3417 |
| language | English |
| publishDate | 2025-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Applied Sciences |
| spelling | doaj-art-13a12585b60d450b823050814e5bc4c42025-08-20T02:48:02ZengMDPI AGApplied Sciences2076-34172025-02-01153165410.3390/app15031654Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask EncodingIrfanullah Memon0Muhammad Ammar Ul Hassan1Jaeyoung Choi2School of Computer Science and Engineering, Soongsil University, 369 Sangdo-Ro, Seoul 06978, Republic of KoreaSchool of Computer Science and Engineering, Soongsil University, 369 Sangdo-Ro, Seoul 06978, Republic of KoreaSchool of Computer Science and Engineering, Soongsil University, 369 Sangdo-Ro, Seoul 06978, Republic of KoreaFew-shot font generation seeks to create high-quality fonts using minimal reference style images, addressing traditional font design’s labor-intensive and time-consuming nature, particularly for languages with large character sets like Chinese and Korean. Existing methods often require multi-stage training or predefined components, which can be time-consuming and limit generalizability. This paper introduces Patch-Font, a novel single-stage method that overcomes the limitations of prior approaches, such as multi-stage training or reliance on predefined components, by integrating a patch-based attention mechanism and a multitask encoder. Patch-Font jointly captures global style elements (e.g., overall font family characteristics) and local style details (e.g., serifs, stroke shapes), ensuring high fidelity to the target style while maintaining computational efficiency. Our approach incorporates triplet margin loss with hard positive/negative mining to disentangle style from content and a style fidelity loss to enhance local style consistency. Experiments on Korean (printed and handwritten) and Chinese fonts demonstrate that Patch-Font outperforms state-of-the-art methods in style accuracy, perceptual quality, and generation speed while generalizing robustly to unseen characters and font styles. By simplifying the font creation process and delivering high-quality results, Patch-Font represents a significant step forward in making font design more accessible and scalable for diverse languages.https://www.mdpi.com/2076-3417/15/3/1654few-shot font generationstyle transferpatch-based attentionmultitask encodingimage-to-image translation |
| spellingShingle | Irfanullah Memon Muhammad Ammar Ul Hassan Jaeyoung Choi Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding Applied Sciences few-shot font generation style transfer patch-based attention multitask encoding image-to-image translation |
| title | Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding |
| title_full | Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding |
| title_fullStr | Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding |
| title_full_unstemmed | Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding |
| title_short | Patch-Font: Enhancing Few-Shot Font Generation with Patch-Based Attention and Multitask Encoding |
| title_sort | patch font enhancing few shot font generation with patch based attention and multitask encoding |
| topic | few-shot font generation style transfer patch-based attention multitask encoding image-to-image translation |
| url | https://www.mdpi.com/2076-3417/15/3/1654 |
| work_keys_str_mv | AT irfanullahmemon patchfontenhancingfewshotfontgenerationwithpatchbasedattentionandmultitaskencoding AT muhammadammarulhassan patchfontenhancingfewshotfontgenerationwithpatchbasedattentionandmultitaskencoding AT jaeyoungchoi patchfontenhancingfewshotfontgenerationwithpatchbasedattentionandmultitaskencoding |